On Thu, 25 Apr 2002, mukesh agrawal wrote:
> A question on the way wait_xact_lmt works... When wait_xact_lmt is
> reached, we drop any additional requests.
Yes.
> If the request was a normal request, then the robot will go back
> to sleep. As expected, we just drop the event.
Correct.
> But if the request is for an embedded object (or if it is a
> redirect request), then we don't increment theExtraLaunchLvl in
> the case where wait_xact_lmt has been reached.
True. We do not increment the extra launch level because we did not
make an [extra] launch.
> I'm wondering if this is intentional, or if it is an oversight.
I think this is intentional, but there are probably cases where you
would want wait_xact_lmt to behave differently. The original design
had a simple goal: prevent memory exhaustion by dropping requests when
the queue gets too long. Under that design, we wanted to get back to
configured request rate as soon as possible. Thus, we left extra
launch level unchanged so that the traffic on the wire picks up to the
configured level as soon as the queue is drained (e.g., when excessive
queuing was caused by a temporary problem).
Would you like wait_xact_lmt to do something else instead?
Thanks,
Alex.
This archive was generated by hypermail 2b29 : Mon Feb 06 2006 - 12:00:22 MST